The Enigmatic Sixth Wave of the Electrocardiogram: the U Wave

The enigmatic sixth wave of the electrocardiogram: The U wave

Andrés Ricardo Pérez Riera1, Celso Ferreira1, Celso Ferreira Filho1, Marcelo Ferreira1, Adriano Meneghini1, Augusto Hiroshi Uchida2, Edgardo Schapachnik3, Sergio Dubner4, Li Zhang5

1For International VCG Investigators, ABC Faculty of Medicine (FMABC), Foundation of ABC (FUABC), Santo André, Brazil 2Heart Institute of the University of São Paulo Medical School, São Paulo, Brazil 3Dr. Cosme Argerich Hospital, Buenos Aires, Argentina 4Clínica and Maternidad Suizo Argentina, Buenos Aires, Argentina 5Main Line Health Heart Center, Jefferson Medical College, Philadelphia, PA, USA Electro-Vectorcardigraphic Section, ABC Medical School, ABC Foundation, Santo André, São Paulo, Brazil

Abstract The U wave is the last, inconstant, smallest, rounded and upward deflection of the electrocardiogram. Controversial in origin, it is sometimes seen following the T wave with the TU junction along the baseline or fused with it and before P of the following cycle on the TP segment. In this review we will study its temporal location related to monophasic action potential, cardiac cycle and heart sounds, polarity, voltage or amplitude, frequency and shape- contour. We will analyze the clinical significance of negative, alternant, prominent U wave,

and the difference between T wave with two peaks (T1–T2) and true U wave. Finally we will analyze the four main hypotheses about the source of U wave: repolarization of the intraventricular conducting system or Purkinje fibers system, delayed repolarization of the papillary muscles, afterpotentials caused by mechanoelectrical hypothesis or mechanoelectrical feedback, and the prolonged repolarization in the cells of the mid-myocardium (“M-cells”). (Cardiol J 2008; 15: 408–421) Key words: U wave, sixth wave of electrocardiogram, source hypothesis

Introduction baseline or fused with it and before P of the following cycle on the TP segment. Usually, U wave has The U wave is the last, inconstant, smallest the same polarity as the T wave. Patients with ne- (£ 1 mm or 11% of voltage of the precedent T wave gative T waves and positive U waves are called with a range of 3% to 24%), rounded and upward de- Type I discordance. Patients with positive T waves flection (except VR, under normal conditions U wave and negative U waves are named Type II discordan- is only negative in VR) of the electrocardiogram ce. From 18,750 consecutively recorded ECGs, (ECG). Controversial in origin, it is sometimes seen 53 patients had type I and 26 type II discordance. following the T wave with the TU junction along the Types I and II were called group A and B respectively.

Address for correspondence: Andrés Ricardo Pérez Riera, MD, Rua Sebastião Afonso, 885-Jd. Miriam 04417-100, Sao Paulo, Brazil, tel: (55 11) 5621-2390, fax: (55 11) 5625-7278/5506-0398, e-mail: [email protected] Received: 23.08.2008 Accepted: 30.08.2008

408 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

Figure 1. At higher rates the TP segment and U wave disappear when the T wave merges with the following P wave.

Patients with negative T and negative U waves paper recorded the shadow. As the heart contracts (concordant polarity) were named as group C. Co- and relaxes repeatedly, Einthoven could record the ronary disease was slightly more common in wave’s pattern of these impulses. The electrocar- Group A (64%) than in Group B (46%, p = 0.174; diograph machine of today looks very different but NS). Coronary artery disease in Group C was extre- works on the same principle. mely common (88%; p < 0.001). Hypertension in We will study in the U wave, the temporal loca- the two discordant groups was similar: Group A tion related to monophasic action potential (AP), car- (60%) versus Group B (58%; p = NS), Group C was diac cycle and heart sounds, polarity, voltage or ampli- significantly higher (88%, p < 0.001). Left ventri- tude, frequency, normal duration and shape contour. cular enlargement (LVE) was 49% in Group A and 58% in Group B (p = NS), but Group C was signifi- Location of U wave cantly higher at 70% (p = 0.038). The authors found U wave needs to be recognized in relationship that the significance of any U wave is not indepen- to the following biological signals. dent of their respective T wave. They propose that the U wave should not be analyzed in isolation, but Monophasic action potential rather with respect to its T wave [1]. The ratio of U wave is coincident with phase 4 of AP. This the U wave and T wave amplitude is relatively con- phase of the AP is associated with diastole of the stant in all leads. The U wave is the last integrant of chamber of the heart (Fig. 2). ventricular repolarization together with ST segment, J wave (Osborn) and T wave. In a healthy person at Cardiac cycle low heart rates the PR, ST and TP segments are at In men under normal conditions, the temporal the same level and form the isoelectric line (Fig. 1). analysis of all phases of cardiac activity shows us In 1896 Einthoven [2], using an improved elec- that the U wave is registered during the protodia- trometer and a correction formula developed inde- stolic period of the cardiac cycle (diastolic isovolu- pendently of Burch, distinguishes only five deflec- metric phase and of fast filling) (Fig. 2). tions on ECG which he names P, Q, R, S, and T. The great master identified the U wave only a few Cardiac sounds years later in 1903, when he used the string galva- The U wave is concomitant to the second (S2) or nometer [3]. He published the first organized pre- third (S3) cardiac sounds. The S2 is produced by closu- sentation of normal and abnormal electrocardio- re of the aortic and pulmonary valves (A2 and P2), at grams recorded with a string galvanometer [4, 5]. the end of ventricular systole, and at the beginning of

In this manuscript, Einthoven described for the first ventricular diastole, S3 sound occurs after S2 (Fig. 3). time the U wave among other waves. The U wave was detected only in ECGs made with the string ga- ECG surface lvanometer. This labeling was used routinely after The distance from the end of T wave until the tracings were made with the galvanometer(adapted apex of U wave is between 90 to 110 ms with ranges from [6]). Willem Einthoven, developed the first of heart rates of 50 to 100 beats/min. The distance electrocardiograph machine. It was a simple string end of T wave/end of U wave is 160 to 230 ms. galvanometer — capable of measuring small changes in the electrical potential as the heart contrac- U wave polarity ted and relaxed. Electrodes were attached to the limbs of the patient. As the string deflected, it ob- In the frontal plane, normal U vector is located structed a beam of light and the photographic around + 60°; thus U wave is positive in II, III and

www.cardiologyjournal.org 409 Cardiology Journal 2008, Vol. 15, No. 5

Figure 2. The U wave is coincident with phase 4 of monophasic action potential. The U wave in surface electrocardiogram is located on the TP interval. It extends from the end of the T wave up to the P wave of the following cycle.

Figure 3. The U wave of electrocardiogram and the contemporary moments of mechanical cycle of the heart and its relationship with the second sound.

VF, and negative in VR and isoelectric in VL. Fre- ECG abnormalities in ca 93% of cases. The main cau- quently the U wave has equal polarity to the prece- ses of negative U wave on ECG are listed below [7]: ding T, i.e. positive where T also is (Fig. 4). — coronary artery disease; In precordial leads, U vector points towards the — hypertension (near 40% of cases); left and the front. Thus, U wave is positive and bet- — valvular heart disease; ter observed in V3 (between V2 and V4) (Fig. 5). — congenital heart disease; — hyperthyroidism; Causes of inverted U wave — primary cardiomyopathy; — without heart disease (ca 7% of cases). A negative U wave is highly specific for the pre- Additionally, a negative U wave is considered sence of heart disease and is associated with other an indirect criterion of LVE.

410 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

dyskinetic region. Certainly, the clinical usefulness of U waves remains underutilized in the spectrum of coronary artery disease. When U-wave changes are the first and only sign of ischemia, they may contribute to a decision regarding the hospital admission of a patient without typical ischemic symptoms [10]. In patients with acute ischemia, U wave inver- X sion could be explained by [11]: — abnormal relaxation pattern during the protodiastolic phase; “protodiastolic shortening” or aftercontraction; — asynchronous segmental early relaxation,which is defined as a localized early segmental outward motion of the left ventricular endocardium during isovolumetric relaxation; — altered left ventricular relaxation rate. Effort-induced U-wave inversion in the precordial leads has long been recognized as a marker of Figure 4. Normal location of U axis in frontal plane. stenosis of the left anterior descending coronary artery, but this pattern is seldom taken into acco- unt [12]. Lone U wave inversion after exercise or exercise-induced inversion of the U-wave is highly predictive of significant coronary artery disease and, more specifically, of disease of the proximal left anterior descending coronary artery (specificity: 93%; sensitivity 23%) [13].

During attack of Prinzmetal angina Transient U wave inversion can be caused either by regional myocardial ischemia or by hypertension. The characteristics of U wave inversion during chest pain attacks in 21 patients with variant angina were compared with those observed in 38 patients with hypertension without apparent ischemic heart disease. Differentiation was possible according to the ECG phase in which U wave inversion appeared. U wave inversion was conside- Figure 5. The U wave is better observed in precordial leads (semi direct leads) when compared to frontal pla- red to be significant if there was a discrete negati- ne leads (indirect leads). Usually the tallest are found in ve deflection of more than 0.05 mV within the TP leads V2, V3 or V4. U wave is normally positive in all segment. U wave inversion proceeded to positive precordial leads. deflection of U wave in patients with hypertension without ischemic heart disease (initial U wave inversion). In contrast, inverted U wave occurred Coronary artery disease after positive U wave deflection during attacks in patients with variant angina (terminal U wave in- Negative U waves, when present, may be of version). When cold pressor test was performed in immense clinical importance. Inverted U wave is patients with variant angina during treatment with a specific electrocardiographic sign of cardiac ische- calcium antagonists, no patient had either anginal mia [8]. It is not only an early noninvasive marker attacks or ischemic ST-segment deviation, but 43% for acute ischemia, but it also disappeared after had transient initial U wave inversion, which was successful revascularization [9]. In patients with followed by positive U wave deflection. U wave in- ischemic heart disease, the U wave vector tends to version can be classified as initial U wave inversion be directed away from the site of the akinetic or and terminal U wave inversion according to the

www.cardiologyjournal.org 411 Cardiology Journal 2008, Vol. 15, No. 5

V3 U

U V4

Figure 6. This electrocardiogram belongs to a 66-year-old man, carrier of Prinzmetal’s variant angina, recorded during a severe vasospastic crisis. The coronary artery shows a noncritical lesion with < 50% of obstruction. Intermittent negative U wave was observed concomitant with ST-segment elevation.

phasic relationship to positive U wave deflection; after a coronary arterial bypass graft procedure is the latter is observed in association with regional associated with a decrease in the QRS amplitude myocardial ischemia. The former seems to be rela- but with no consistent changes in T wave polarity. ted to elevated blood pressure rather than to myocardial ischemia [14] (Fig. 6). Value of intracoronary electrocardiography Negative U waves in the precordial leads on Value of intracoronary electrocardiography admission can be seen in about 20% of the cases of (IC-ECG) [17] is a sensitive tool in detecting myo- acute anterior myocardial infarctions, mainly in tho- cardial ischemia. It is a more sensitive method than se smaller, with less ST segment elevation, better surface ECG to detect electrical changes during collateral circulation, and larger amount of stunned percutaneous transluminal coronary angioplasty viable myocardium [15]. In this circumstance, (PTCA). It also provides direct monitoring of ST-T U wave inversion had no predictive value in locali- segment, QTc intervals, and U-wave genesis during zing the diseased artery. balloon inflation. PTCA of left anterior descending It is suggested that within the appropriate cli- artery (LAD) was performed in 14 patients using the nical context, an isolated U-wave inversion may standard balloons and in 11 patients using the perfu- precede the typical electrocardiographic changes of sion balloons. Patients with perfusion balloon angio- an acute myocardial infarction by several hours [16]. plasty had: less ST-T elevation, less ischemia, less The change from a negative to an upright U wave QTc-shortening intervals and less positive U waves.

412 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

U-wave changes on the IC-ECG during anterior Valvular heart disease or inferoposterior myocardial ischemia were corre- lated with the U-wave changes in the precordial le- An inverted U wave may represent valvular ads of the body surface ECG in 28 patients who un- heart disease (mainly aortic and mitral regurgita- derwent PTCA of the LAD group (17 patients) or left tion) [21]. circumflex (LCx) (LCx group; 11 patients). An inverted U wave may represent left ventri- The IC-ECG was recorded simultaneously with cular volume overload [22]. Negative U wave in left the body surface multiple precordial leads at the leads is considered an indirect criterion of LVE. Both baseline and during PTCA. the precordial T-wave imbalance and the U-wave The amplitude of the U-wave on the IC-ECG negativity are common findings in early LVE [23]. was measured quantitatively, and U-wave changes Right ventricular enlargement: negative from baseline to PTCA were assessed qualitative- U wave in right precordial leads is observed in right ly on the body surface ECG. ventricular enlargement. The U wave vector is di- Three different patterns of U-wave changes rected opposite to the QRS axis in the horizontal were distinguishable on the IC-ECG from baseline plane in patients with both left and right ventricu- to angioplasty: change to positivity, no change and lar enlargement [7]. change to negativity. The incidence of each pattern was similar in the LAD and LCx groups. The IC-ECG U wave alternans was more sensitive for detecting U-wave changes The U wave alternans is an electrocardiogra- during PTCA than body surface precordial ECG. phic sign of left ventricular failure and increased When compared to the IC-ECG, concordant U-wave ventricular irritability [24, 25]. changes occurred in the surface precordial ECG in 67% (8/12) of the LAD group with accompanying U-wave voltage or U-wave amplitude epicardial U-wave changes, and discordant changes It is normally always lower than 50% of the in 33% (3/9) of the LCx group with epicardial width of the preceding T and generally between U-wave changes [18]. 3% and 24% of it. Usually it does not exceed 1 mm, being in average of 0.33 mm. If it reaches 1.5 mm Hypertension or more, it is considered prominent or high, howe- ver, there may be normal U waves of up to 2 mm

U inversion is observed in nearly 40% of cases (0.2 mV) in II and from V2 to V4. The tallest positi- of high blood pressure patients. Transient U wave ve U wave is usually observed in the area of leads inversion can be caused by an elevation of syste- V2 to V4. U-wave voltage is inversely proportional mic blood pressure. Negative U wave in left pre- to heart rate (RR interval). cordial leads is considered and indirect signal of LVE. The deepest negative U wave is usually ob- Causes of prominent U waves served in the area of leads V5 to V6 [19]. — Bradycardia. U wave voltage is strongly rate The ECGs of 297 cases of hypertension were dependent (inversely proportional). U wave is divided into 6 groups on the basis of the relation- observed better during bradycardia. When he- ship between the polarity of the T wave and the art rate (HR) is £ 65 bpm, U waves are visible U wave. Both waves were positive in all precordial in 90% of cases. When HR is between 80 bpm leads in 48.1% of the cases. Negative U waves were and 95 bpm U waves are visible in 65% of ca- found in 21.8% of the cases and these were predo- ses. When HR is > 95 bpm U waves are visi- minantly in the leads with negative T waves. A ne- ble in 25% of cases (Fig. 7, 8). Bimodal T wa- gative T wave in V5 and V6 was accompanied most ves with hump-like morphology represent dif- frequently by a negative, less frequently by an iso- ferent levels of interruption of the descending electric, and least frequently by a positive, U wave. slope of the T wave, called T2 instead of An inverted U wave in the presence of an upright U wave. Bimodal or notched T waves may be T wave was found in only 2.8% of the cases. A chan- distinguished from the T-U interval: the second ge from a negative to a positive or isoelectric U apex of bimodal T wave (T2) is at a distance wave was observed after slowingof the heart rate, from the first one (T1) < 150 ms; the T1-U in- a drop in blood pressure, and nitroglycerin admi- terval is > 150 ms. The second apex of bimo- nistration [20]. dal T wave (T2) is at a distance < 150 ms from

www.cardiologyjournal.org 413 Cardiology Journal 2008, Vol. 15, No. 5

Figure 7. Percentage of U wave visualization related with heart rate (HR).

Figure 8. Electrocardiogram differentiation between T2 (bimodal T-wave) and U wave (HR).

the first module (T1): The T1-U interval is al- reversal in the relative voltage of the T and ways > 150 ms [26]. U wave increases its vol- U waves, QTc interval prolongation, tendency tage or appears during slow rates or after pau- to torsade de pointes (TdP) atypical tachycar- ses [27], as in the cases of long QT interval. dia and digitalis action enhancement. Hypoka- In this case we observe a bifid T wave and lemia enhances the tachyarrhythmias produced a U wave only with slow HR. by digitalis toxicity. — Early repolarization variant (ERV). Becau- — Hypomagnesemia. Prominent and alternant se bradycardia U waves are frequent, in ERV U wave with ventricular irritability was descri-

they are best observed in the V3 lead. U waves bed in patients with low serum magnesium are frequent when sinus bradycardia is present. (< 1.8 mmol/L) [25]. — Hypokalemia. Hypokalemia is associated — Hypocalcemia. with flattening of the T wave and the appearan- — Hypothermia. ce of a prominent U wave. A pathological — Class III antiarrhythmic drugs (amiodarone, “U” wave as seen with hypokalemia is the conse- dofetilide, sotalol). quence of electrical interaction among ventri- — Class IA (quinidine, disopyramide, and proca- cular myocardial layers at AP phase 3 of which inamide). Quinidine effect is characterized by repolarization slows [28]. The main ECG featu- discrete QRS prolongation; only with an extre- res of hypokalemia are: PR interval prolonga- me quinidine effect, significant QT/QTc inte- tion, gradual ST segment depression ≥ 0.5 mm rval prolongation, depressed, widened, not-

in II or from V1 to V3, decrease of T wave ampli- ched, and inverted T waves, prominent U wa- tude (flat T wave), possible T wave inversion, ves and TdP tendency. prominent U wave (U wave > 0.5 mm in II or — Digitalis effect or digitalis action. The ear-

> 1 mm in V3) secondary to prolongation of the liest modification of digitalis effect on ECG or recovery phase of cardiac AP, characteristic “digitalis action” are: PR interval prolongation,

414 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

ST segment: shortening and superior convexity QTU seems to be characteristically associated (“in spoon”) by shortening of phases 2 and 3 of with TU alternans. Dispersion of repolarization AP, shortening QT and QTc intervals (main cau- may underlie the increased ventricular electri- se of acquired short QT interval), flattening with cal instability in these cases [32]. Andersen- apiculate T wave of terminal portion in 10% of Tawil syndrome (ATS) is a channelopathy af- cases, possible symmetrical inversion of T wave fecting inward rectifier potassium I(K1) with (pseudo-ischemic T wave) and prominent U wave. QT prolongation, large U waves, and frequent — Phenothiazines. The electrophysiological ventricular tachycardia (VT) [33]. properties of phenothiazines are comparable to — Left circumflex-related myocardial infarc- those of the Class Ia antiarrhythmic quinidine. tion. ECG criteria of strictly posterior myocar- Numerous ECG aberrations may be induced by dial infarction with the left circumflex corona- these agents, including changes in the morpho- ry artery as an infarct-related coronary artery logy of the T wave, prolongation of the QT inte- apply at less than 6 hours or at 24 hours since rval, and accentuation of the U wave. Even with the onset of the symptoms. 1) ST depression standard clinical dosages (100–400 mg/d.), thio- ≥ 0.1 mV in two consecutive chest leads; 2) pro-

ridazine causes minimal prolongation of the QT minent positive U wave ≥ 0.1 mV in leads V2 or V3;

interval, reduction of T wave amplitude, and pro- 3) T/U ratio in leads V2 or V3 £ 4. Considering minent U waves in nearly 50% of patients. two of the above criteria as positive, the sensi- — Forced inspiration. tivity was 71.9%, the specificity 97.0%, and the — Post-exercise. diagnostic accuracy 88.8% (Fig. 9) [34]. — Mitral valve prolapse. — Left ventricular enlargement. U wave duration — Alterations of the central nervous system It is about 170 ± 30 ms in normal adults. with endocranial hypertension. In patients with cerebrovascular event T and U waves are U wave contour augmented, consequently the T/U ratio is not The normal U wave has an asymmetric shape altered. with rapid ascending limb and more slow descen- — Cardiomyopathies: i.e. Hypertrophic cardio- ding limb (just the opposite of the normal T wave). myopathy with solitary hypertrophy of papillary muscle [29]. U wave constance — Acquired complete atrioventricualr block. Frequently absent (50% to 75%) [35]; occasio- Torsades de pointes during acquired complete nally hard to distinguish from the preceding T wave. atrioventricular block is rare. The predictors of Better observed during bradycardia and sometimes ventricular arrhythmias during acquired comple- related to TdP. te atrioventricular block are the presence of prolonged QTc/JTc intervals, pathologic U wave Terms definitions and T-U complex, prolonged Tpeak-Tend inte- QT interval. Known as electric “systole”. It rval, and LQT2-like QT morphology [30]. corresponds to ventricular depolarization and repo- — Congenital long QT syndrome. Augmentation larization. Measurement should be conducted in VL and greater degree of merging of the T and U so as not to include the U wave. The end of the T waves and QTc interval prolongation are changes wave is the intersection of a tangent to the steepest alerting about the possibility of congenital long QT slope of the last limb of the T wave and the baseli-

syndrome, specifically genotype 2 or 1 [31]. ne, in lead II or V5 [36]. T or U wave alternans in association with long Measurement of the QT interval must be con- QTU and TdP is uncommon and its mecha- ducted at a double velocity of 50 mm/s and with the nism(s) is unknown. 1) TU alternans may be 12 leads vertically aligned, with the aim of analy- due to 2:1 propagation of an early after depola- zing one beat simultaneously from the first deflec- rization (EAD) or to alternation of the recove- tion of the QRS complex until the point of return of ry kinetics of a repolarization current; 2) The the T wave to the baseline or in the lowest point constant occurrence of EAD in relation to pha- between the T and U waves [37]. se 0 in spite of alternation of plateau duration Q-aT interval. From the onset of Q up to the suggests an ionic mechanism synchronized to apex of T. depolarization; 3) Tachycardia dependent TdP Q-aTc interval. Interval from the onset of Q in clinical and experimental examples of long up to the apex of T, divided by the root square of RR.

www.cardiologyjournal.org 415 Cardiology Journal 2008, Vol. 15, No. 5

Figure 9. This electrocardiogram was obtained in a 75-year-old asymptomatic female a history of myocardial infarction. This tracing shows sinus bradycardia at 55 beats/min. The abnormal Q waves and nonspecific ST-T changes in leads II, III and VF, and prominent R waves in leads V1 and V2 are consistent with prior inferolateral infarction (new topographic myocardial infarction classification). Coronary artery disease should also be coded. Prominent positive U wave ≥ 0.1 mV in leads V2 and V3.

Q-oT. Interval from the onset of Q up to the Horizontal plane. The U loop is directed to onset of T. front and the left between V2 and V4. Q-aU. Interval from the onset of Q up to the apex of U. Digitalis frequently cause separation The source of U wave between the end of T wave from the U wave be- The source of the U wave is uncertain. Several cause digitalis shortens the QT interval. hypotheses for the origin of the U wave in ECGs have QT + U interval or Q (T + U) interval. This been proposed. A satisfactory explanation of its ori- term is employed in cases of marked QT interval gin is still outstanding. Various explanations have been lengthening (congenital long QT, severe bradycar- offered for their origin, but none is universally accep- dia, hypokalemia and hypocalcemia, fad diets, con- ted. The surface potential at any given electrode lo- trast injection into coronary artery, administration cation is the net result of simultaneously acting and of class IA and III antiarrhtymics agents, post-re- variously directed electrical forces in the myocardium. suscitation, neurogenic, organophosphorus intoxi- Four hypotheses are being entertained about cation) when the QT interval is prolonged more than the underlying mechanism: 100 ms (≥ 125%) at heart rates within 50 to — tardy repolarization of several endocardial 100 beats/min, T occupies the territory of the structures: U wave and masks it [38]. — repolarization of the intraventricular con-

When the T wave has two peaks (T1–T2), the duction system or Purkinje fibers system. distance between the first peak and the second one This is the Hoffmand and Crannefield hy- is always < 0.15 s (150 ms) and the distance betwe- pothesis [39]; en the first T peak and the U peak wave is > 0.15 s — delayed repolarization of the papillary (150 ms) [26]. muscles named by Bufalari and Furbet- ta “the syndrome of the papillary musc- The normal U wave les” [40]; loop vectorcardiogram — afterpotentials caused by mechanical forces Frontal plane. The U loop is directed to +60° in the ventricular wall. Electro-mechanic near the II lead. Normally, it is always projected into coupling, mechanoelectrical hypothesis or the positive segment of the II lead. U wave may be mechanoelectrical feedback; inscribed in a counterclockwise, in a clockwise, or — prolonged repolarization in the cells of the an 8 configuration. mid-myocardium (“M-cells”).

416 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

Tardy repolarization of several — left papillary muscle syndrome: negative endocardial structures U wave in left leads I, VL, V5–V6, because spa- Repolarization of the intraventricular con- tial U vector is directed to the front and right. ducting system or Purkinje fibers system. This Observed in anterior myocardial infarction, hy- is the Hoffmand and Crannefield hypothesis [39]. pertension and aortic valvular disease; It has been suggested that the U wave may reflect — right papillary muscle syndrome: negative repolarization of the ventricular Purkinje system, U wave is observed in III, sometimes in VF, since the AP duration in these fibers is the longest and right precordial leads. Observed in right for any fibers found in the heart. In an experimen- ventricular enlargement and congenital heart tal canine model, Watanabe conducted a compari- disease; son of the AP of Purkinje and ventricular muscle — biventricular papillary muscle syndrome: ne- fibers under conditions accentuating the U wave. gative U wave in all precordial leads, II and VL. The author concludes that the U wave represents Biventricular enlargement (strain of both ven- Purkinje fibers repolarization. The author observed tricles). a good temporal correlation between phase 3 repo- The phenomenon of solitary papillary muscle larization in Purkinje fibers and the electrocardio- hypertrophy in hypertrophic cardiomyopathy graphic U wave [41]. (HCM) is very rare. Giant negative T and U waves Against this theory we have the following ar- are two common ECG phenomena in HCM and have guments: been attributed to hypertrophy of the posterior pa- — the small mass of Purkinje fibers, in relation pillary muscle. Solitary hypertrophy of the anterior to the very large mass of ventricle, may be in- papillary muscle might be a new echo-electrocar- sufficient to affect the ECG and generate the diographic syndrome related with the Furbetta hy- U wave [42]; pothesis [29]. — amphibian hearts have no Purkinje fibers but Apical hypertrophy, especially of the posterior do show U waves [43]; papillary muscle, may play an important role in the — the timing of the U wave during ventricular re- pathogenesis of Giant negative T and negative laxation and the links between U wave and U waves in HCM [46]. mechanical events favor the mechanoelectrical hypothesis of U wave genesis [44]; Electro-mechanic coupling, — the timing of the U wave apex is dependent on mechanoelectrical hypothesis, the duration of ventricular repolarization but mechanoelectrical feedback not on the QRS duration. This observation is Controversies persist regarding whether the better explained by the ventricular relaxation U wave is a purely electrical or mechanoelectrical than by the Purkinje fiber repolarization the- phenomenon. The characteristics of U wave are not ory of U wave genesis; compatible with the Purkinje or ventricular muscle — in the presence of complete right bundle branch repolarization hypotheses. The timing of the U wave block, the timing of the U wave correlates better during ventricular relaxation and the links between with the presence of right ventricular enlarge- U wave and mechanical events favor the mechano- ment than with the dromotropic disturbance [45]. electrical hypothesis of U wave genesis [44]. Schimpf Delayed repolarization of the papillary et al. [47]. used echocardiographic measurements muscles is the denomination created a long time to discriminate between the hypotheses for the ago (1956) by Drs. D. Furbetta and A. Bufalari [40]. origin of the U wave. Echocardiography and ECG They identify ”the papilary muscle syndrome” and were performed in 5 SQTS patients from 2 unrela- postulate that the U wave represents repolarization ted families with a history of sudden cardiac death of the papillary muscle and neighboring structures. and 5 age-matched and gender-matched control sub- The common factor underlying the varied cardiac jects. In SQTS patients, the end of the T wave pre- pathology is regarded to be ischemia, ”strain”, or ceded aortic valve closure by 111 ± 30 ms vs. –12 ± other functional derangement of the papillary musc- ± 11 ms in control subjects. The interval from aor- les in the right or left ventricles. The authors belie- tic valve closure to the beginning of the U wave was ve that various abnormalities of the papillary musc- 8 ± 4 ms in patients and 15 ± 11 ms in control sub- les, whether anatomic or functional, are detectable jects. Thus, the inscription of the U wave in SQTS by modifications on U waves and T-U segment. patients coincided with aortic valve closure and iso- They described three different vectorial pat- volumic relaxation, supporting the hypothesis that terns: the U wave is related to mechanical stretch. There

www.cardiologyjournal.org 417 Cardiology Journal 2008, Vol. 15, No. 5 is a significant dissociation between the ventricu- gree to which the electrical forces in the heart are lar repolarization and the end of mechanical systo- opposing each other. This mechanism plays a ma- le in SQTS patients. Coincidence of the U wave with jor role in the formation of the ECG waves. Ritse- termination of mechanical systole provides support ma van Eck et al. [52] used a digital computer mo- for the mechanoelectrical hypothesis for the origin del of the left ventricle to study the effect of source of the U wave [48]. There is a reason to believe that locations on cancellation during the T and U waves. the U wave is produced by potentials elicited by the The model represents an anatomically stylized stretching of ventricular muscle during the period cross-sectional slice of the left ventricle, containing of rapid blood inflow into the ventricles [43]. Post- 1961 hexagonal cells in a single layer. The model potentials or afterpotentials with triggered activity employs a multi-layered segment of the myocar- mechanism are the presumable genesis. Afterpo- dium. An AP is assigned to each cell. The timing of tentials occur in stretched cardiac fibers and the the APs follows a simulated excitation sequence. subendocardial tissue is subjected to a greater stretch The potential differences between the APs of adja- than the subepicardial muscle. Lepeschkin [49] cent cells produce time-varying electrical sources, has stated that the U wave results from potential each of which contributes to the potential in an ar- differences between the ventricular muscle with bitrary point P on the body proportionally to its own, larger negative afterpotentials and that with smal- location-dependent, transfer function (lead vector). ler afterpotentials, with the latter positive relative The ECG at P is the sum of all potential contribu- to the former. Di Bernardo and Murray [50] pro- tions. The potential differences between the APs posed a simple method to model repolarization in produce time-varying electrical sources. Each so- the left ventricle (LV) and the corresponding urce contributes to the potentials in an arbitrary T waves on the surface ECG. The authors modeled point P of the body. The strength of this contribu- the cardiac cell APs in the LV with differences in tion is determined by a specific coefficient, the “lead only the duration of the plateau phase. Using publi- vector”, linking P to the source. The ECG recor- shed experimental data on the epicardial and endo- ded at P is calculated as the sum of all potential cardial repolarization sequences, for each point on contributions. For each time point in the ECG at P, the LV surface the authors set a different AP repo- the contribution of each cell is mapped back onto larization starting time, determined by the duration the slice. Adjacent cells with equal contributions of the plateau phase. The surface source model was form iso-source strings, together forming iso-sour- used to compute potentials on the surface of the ce maps. The T-U wave as observed in P will be torso, generated by repolarization of the LV. Both the sum of positive and negative contributions from the torso and the LV had homogeneous and isotro- the iso-source distributions as they change with pic conductivity. They simulated T waves on the time. The iso-source maps for an anteriorly loca- 12-lead ECG and compared their results with me- ted observation point P at 4.2 cm from the epicar- asured T waves from five normal subjects. The dial surface show a continuous interplay of positive orientation and shape in each lead were reprodu- and negative contributions. During the near-zero ST ced. In each lead the authors computed the root segment, cancellation varies between 80% and mean square error between simulated and measu- 100%. In the ascending limb of the T wave, positi- red T waves. The average error across the 12 le- ve contributions substantially increase, giving ads was small, with a mean value of 0.11 mV across a decrease in cancellation to about 40%. At the end all the subjects. The authors concluded that repo- of the T wave (with almost zero amplitude), the larization of the LV can be modeled independently positive contributions are only slightly reduced as of the depolarization sequence and AP duration gra- compared with those at peak T, but greatly incre- dients. This method is an easy and powerful tool to ased negative contributions cancel them out. This describe the ECG features of repolarization. The is contrary to the generally held view that the end same author using a computer model of LV repola- of T signifies the end of the repolarization process. rization demonstrated that a delaying repolarization The manifest shape of the T and U waves is the in different regions of the heart cannot explain the result of complex interactions of varying and often U-wave. The presence of after-potentials on the car- largely canceling contributions. The iso-source diac AP does explain the U-wave polarity and other maps are helpful to understand the genesis of the characteristic U-wave features. The authors also T and U waves [52]. The repolarization waves con- show that abnormal after-potential timing corre- structed in this way reproduce the natural aspects sponds with abnormal U-wave inversion [51]. Can- of a T wave followed by a U wave. The creation of cellation in electrocardiology is defined as the de- a U wave is conditional on small voltage differences

418 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

Figure 10. The thickness of the ventricular wall is formed by three functional layers, with different action potential. In the depth of the middle layer we find M cells, which have a subpopulation of cells with a great conduction velocity and electrophysiological properties of their own, which are very relevant in the pathophysiology of long and short QT syndromes.

between the tail ends of the APs. No fundamental differentiated electrophysiological and pharmacolo- demarcation exists between U wave and the pre- gical features. The authors from the Masonic Me- ceding T wave. The morphology of the T-U wave dical Research Laboratory suggest that M-cells, is dependent on the geometrical position of P with more abundant in mass and having a prolonged re- respect to the myocardium. T and U form a conti- polarization time comparable to Purkinje cells, may nuum. Together they are the result of one and the be responsible for the pathophysiologic recording same process of repolarization of the ventricular of the U wave in the presence of acquired or con- myocardium. This has implications for the measu- genital long QT interval. The discovery of M-cells rement of QT duration and for safety testing of drug- and their electrophysiology has established the cel- induced QT prolongation [53]. Depolli et al. [54] lular basis for repolarization and has contributed to used a spatial model of a left ventricle constructed our knowledge of U-wave genesis [55]. from 12 layers composed of cubic cells. Each cell is A second component of an interrupted T wave assigned its own time-dependent AP with its own is more likely to be, and argue for use of the term contribution to the electrical potential at arbitrary T2 in place of U to describe this event [56]. points where ECGs are measured. Simulated ECGs Postextrasystolic U wave augmentation show that U waves can be generated using various (a marked increment in U wave amplitude after pre- combinations of APs across the different layers of mature ventricular contractions is an adverse progno- the ventricular wall. The authors conclude that the stic sign in “pause-dependent long QT syndrome”). U wave can be generated in the presence of strong Postextrasystolic T wave changes are common intercellular coupling. Myocardial layers with pro- and lack predictive value. longed action potentials, like M cells, are not ne- Postextrasystolic U wave changes may be cessarily needed for U-wave genesis. a specific marker of a tendency to the develo- pment of spontaneous ventricular arrhythmias Theory of origin in M-cells observed in (Fig. 10, 11) [57]. acquired or congenital long QT syndrome Besides the three basic types of cells in the Final considerations ventricular myocardium: epicardial, mesocardial and endocardial, there is a cellular subpopulation called Although the sixth wave of the ECG remains “M-cells”, located in the midmyocardium with very a mystery regarding its genesis, new investigations

www.cardiologyjournal.org 419 Cardiology Journal 2008, Vol. 15, No. 5

Figure 11. Characteristics of action potential of M cells. It is a fast and non-automatic fiber; therefore, we could say that it is a mixture between Purkinje cells and contractile ventricular myocardium. It is very sensitive to bradycardia and class III antiarrhythmic drugs, such as amiodarone and sotalol.

indicate the electro-mechanical theory is true, 9. Girish MP, Gupta MD, Mukhopadhyay S, Yusuf J, Sunil Roy TN, except in the cases of long QT interval, where the Trehan V. U wave: An important noninvasive electrocardio- M cell would have a decisive role. As we can see, graphic diagnostic marker. Indian Pacing Electrophysiol J, 2005; 5: 63–65. the clinical observation of this “neglected” small 10. Correale E, Battista R, Ricciardiello V, Martone A. The negative wave has a great clinical significance in very diffe- U wave: A pathogenetic enigma but a useful often overlooked rent pathologies. We believe it is necessary to car- bedside diagnostic and prognostic clue in ischemic heart dis- ry out more experimental and clinical research to ease. Clin Cardiol, 2004; 27: 674–677. provide a basis for the last “enigma” of the ECG. 11. Vanoverschelde JL, Wijns W, Michel X, Cosyns J, Detry JM. Asynchronous (segmental early) relaxation impairs left ventricular filling in patients with coronary artery disease and normal Acknowledgements systolic function. J Am Coll Cardiol, 1991; 18: 1251–1258. 12. Costantini M, Capone S, Tondo A, Oreto G. Is exercise-induced The authors do not report any conflict of inte- U-wave inversion predictive of proximal left anterior descending rest regarding this work. coronary artery disease? J Electrocardiol, 2008; 41: 99–101. 13. Gerson MC, Phillips JF, Morris SN, McHenry PL. Exercise- References -induced U-wave inversion as a marker of stenosis of the left anterior descending coronary artery. Circulation, 1979; 60: 1014–1020. 1. Reinig MG, Harizi R, Spodick DH. Electrocardiographic T- and 14. Miwa K, Miyagi Y, Fujita M, Fujiki A, Sasayama S. Transient U-wave discordance. Ann Noninvasive Electrocardiol, 2005; 10: terminal U wave inversion as a more specific marker for myo- 41–46. cardial ischemia. Am Heart J, 1993; 125: 981–986. 2. Einthoven W. Ueber die Form des menschlichen electrocardio- 15. Tamura A, Watanabe T, Nagase K, Mikuriya Y, Nasu M. Signif- gramms. Arch Ges Physiol, 1895; 60: 101–123. icance of negative U waves in the precordial leads during 3. Snellen HA. Willem Einthoven (1860–1927). Father of electro- anterior wall acute myocardial infarction. Am J Cardiol, 1997; cardiography. Life and work, ancestors and contemporaries. Klu- 79: 897–900. wer Academic Publishers, Dordrecht, 1994. 16. Duke M. Isolated U wave-inversion in acute myocardial infarc- 4. Einthoven W. Die galvanometrische registrierung des menschli- tion. Cardiology, 1975; 60: 220–225. chen Electrokardiogram. Pfluger’s Arch, 1903; 99: 472–480. 17. Safi AM, Kwan T, Feit A, Gonzalez J, Stein RA. Use of intracoronary 5. Einthoven W. Le telecardiogramme. Arch Int Physiol, 1906; 4: electrocardiography for detecting ST-T, QTc, and U wave changes 132–164. during coronary balloon angioplasty. Heart Dis, 2001; 3: 73–76. 6. Hurst JW. Ventricular electrocardiography. Gower Medical Pub- 18. Kataoka H, Yano S, Tamura A, Mikuriya Y. Projection of epicar- lishing, New York 1991: 5–26. dial U-wave change to surface precordial electrocardiogram in 7. Kishida H, Cole JS, Surawicz B. Negative U wave: A highly coronary artery disease. J Cardiol, 1996; 27: 247–254. specific but poorly understood sign of heart disease. Am J Cardiol, 19. Gregory SA, Akutsu Y, Perlstein TS, Yasuda T, Yurchak PM. 1982; 49: 2030–2036. Inverted U waves. Am J Med, 2006; 119: 746–747. 8. Sovari AA, Farokhi F, Kocheril AG. Inverted U wave, a specific 20. Bellet S, Kemp Rl, Surawicz B. Polarity and amplitude of the U electrocardiographic sign of cardiac ischemia. Am J Emerg Med, wave of the electrocardiogram in relation to that of the T wave. 2007; 25: 235–237. Circulation, 1957; 15: 90–97.

420 www.cardiologyjournal.org Andrés Ricardo Pérez Riera et al., The enigmatic sixth wave of the electrocardiogram: The U wave

21. Fisch C. The clinical electrocardiogram: Sensitivity and specificity. 40. Bufalari A, Furbetta D, Santucci F, Solinas P. Abnormality of the ACC Curr J Rev, 1997; 6: 71–75. U wave and of the T-U segment of the electrocardiogram; the syn- 22. Conrath CE, Opthof T. The patient U wave. Cardiovasc Res, drome of the papillary muscles. Circulation, 1956; 14: 1129–1137. 2005; 67: 184–186. 41. Watanabe Y. Purkinje repolarization as a possible cause of the 23. Marriott Henry JL. Challenging ECGs. Section VII: Myocar- U wave in the electrocardiogram. Circulation, 1975; 51: 1030–1037. dial ischemia and infarction. Hanley & Belfus Inc., Philadel- 42. Surawicz B. Electrophysiologic substrate of torsade de pointes: phia 2002: 395. Dispersion of repolarization or early afterdepolarization? J Am 24. Ever KM. U wave alternans: an electrocardiographic sign of left Coll Cardiol, 1989; 14: 172–184. ventricular failure. Am Heart J, 1974; 87: 41–45. 43. Lepeschkin E. Advances in electrocardiography. In: Schlant RC, 25. Bashour T, Rios JC, Gorman PA. U wave alternans and increased Hurst JW eds. Grune & Stratton, New York 1972: 431–447. ventricular irritability. Chest, 1973; 64: 377–379. 44. Surawicz B. U wave: Facts, hypotheses, misconceptions, and 26. Lepeschkin, E. The U wave of the electrocardiogram. Mod Con- misnomers. J Cardiovasc Electrophysiol, 1998; 9: 1117–1128. cepts Cardiovasc Dis, 1969; 38: 39. 45. Ferrero C, Maeder J. Right branch block: Diagnosis of the right 27. Roden DM, Spooner PM. Inherited long QT syndromes: a para- ventricular hypertrophy by the chronology of U the wave. Sch- digm for understanding arrhythmogenesis. J Cardiovasc Elec- weiz Med Wochenschr, 1970; 100: 190–192. trophysiol, 1999; 10: 1664–1683. 46. Hasegawa K, Sawayama T, Nezuo S, Tadaoka S, Kawahara Y, 28. Hlaing T, DiMino T, Kowey PR, Yan GX. ECG repolarization Inoue S. Possible pathogenesis of giant negative T and negative waves: Their genesis and clinical implications. Ann Noninvasive U waves in hypertrophic cardiomyopathy: A report of two cases. Electrocardiol, 2005; 10: 211–223. J Cardiol, 1992; 22: 271–279. 29. Ker J. Solitary papillary muscle hypertrophy: A new echo-elec- 47. Schimpf R, Antzelevitch C, Haghi D et al. Electromechanical trocardiographic syndrome? A case report. Angiology, 2007; 58: coupling in patients with the short QT syndrome: Further in- 502–503. sights into the mechanoelectrical hypothesis of the U wave. 30. Bozkaya YT, Eroglu Z, Kayikçioglu M et al. Repolarization cha- Heart Rhythm, 2008; 5: 241–244. racteristics and incidence of torsades de pointes in patients with 48. Chen PS, Lin SF. Electromechanical coupling in patients with acquired complete atrioventricular block. Anadolu Kardiyol the short QT syndrome: further insights into the mechanoelec- Derg, 2007; 7: 98–100. trical hypothesis of the U wave. Heart Rhythm, 2008; 5: 1091 31. Batchvarov VN, Bajpai Am Behr E. Prolongation of the QT in- (author reply 1091–1092). terval and post-extrasystolic Augmentation of the TU-wave dur- 49. Lepeschkin E. Genesis of the U wave. Circulation, 1957; 15: 77–81. ing emotional stress. Indian Pacing Electrophysiol J, 2008; 8: 50. di Bernardo D, Murray A. Computer model for study of cardiac 146–148. repolarization. J Cardiovasc Electrophysiol, 2000; 11: 895–899. 32. Habbab MA, el-Sherif N. TU alternans, long QTU, and torsade 51. di Bernardo D, Murray A. Origin on the electrocardiogram of de pointes: clinical and experimental observations. Pacing Clin U-waves and abnormal U-wave inversion. Cardiovasc Res, 2002; Electrophysiol, 1992; 15: 916–931. 53: 202–208. 33. Morita H, Zipes DP, Morita ST, Wu J. Mechanism of U wave and 52. Ritsema van Eck HJ, Kors JA, van Herpen G. Dispersion of polymorphic ventricular tachycardia in a canine tissue model of repolarization, myocardial iso-source maps, and the electrocar- Andersen-Tawil syndrome. Cardiovasc Res, 2007; 75: 510–518. diographic T and U waves. J Electrocardiol, 2006; 39: 96–100. 34. Kanemoto N, Wang Y, Fukushi H et al. Electrocardiographic 53. Ritsema van Eck HJ, Kors JA, van Herpen G. The U wave in the characteristics of patients with left circumflex-related myocar- electrocardiogram: a solution for a 100-year-old riddle. Cardio- dial infarction in the acute phase without tented T waves or vasc Res, 2005; 67: 256–262. definite ST elevation. J Cardiol, 1995; 26: 149–158. 54. Depolli M, Avbelj V, Trobec R. Computer-simulated alternative 35. Gammon M. Comparative measurements and statistical studies modes of U-wave genesis. J Cardiovasc Electrophysiol, 2008; on the incidence of U-waves in the electrocardiogram. Schweiz 19: 84–89. Z Gesch, 1954; 2: 82–101. 55. Antzelevitch C, Sicouri S. Clinical relevance of cardiac arrhyth- 36. Postema PG, De Jong JS, Van der Bilt IA, Wilde AA. Accurate mias generated by afterdepolarisations. Role of M cells in the electrocardiographic assessment of the QT interval: Teach the generation of U waves, triggered activity and torsades de tangent. Heart Rhythm, 2008; 5: 1015–1018. pointes. J Am Coll Cardiol, 1994; 23: 259–277. 37. Lepeschkin E, Surawicz B. The measurement of the Q-T 56. Yan GX, Antzelevitch C. Cellular basis for the normal T wave interval of the electrocardiogram. Circulation, 1952; 6: 378–388. and the electrocardiographic manifestations of the long-QT syn- 38. Surawicz B, Knilans TK. Chou S. Electrocardiography in clinical drome. Circulation, 1998; 98: 1928–1936. practice. Chapter 24. 6th Ed. Saunders Elservier, Philadelphia 57. Viskin S, Heller K, Barron HV et al. Postextrasystolic U wave 2008: 578. augmentation, a new marker of increased arrhythmic risk in 39. Hoffmand BF, Cranefield PF. Electrophysiology of the heart. patients without the long QT syndrome. J Am Coll Cardiol, 1996; Mac Graw-Hill, New York 1960: 201–202. 28: 1746–1752.